40 points | by tanelpoder 3 hours ago ago
5 comments
Postgres can do so much. I see people choose Kafka and SQS for things that Graphile Worker could do all day long.
Yo! Author here, I’ll be around if anyone’s got questions!
Did you test with fillfactor < 100 on the queue table? With HOT updates, status changes can reuse dead space without creating new index entries, which seems like it could significantly delay the onset of the death spiral?
It would be nice if this ad at least explained a little bit of the technical side of the solution.
It sounds vaguely like InnoDB’s concurrency control solution which uses tokens [0] as a unit of maximum work a query can perform.
0: https://dev.mysql.com/doc/refman/8.4/en/innodb-performance-t...
Postgres can do so much. I see people choose Kafka and SQS for things that Graphile Worker could do all day long.
Yo! Author here, I’ll be around if anyone’s got questions!
Did you test with fillfactor < 100 on the queue table? With HOT updates, status changes can reuse dead space without creating new index entries, which seems like it could significantly delay the onset of the death spiral?
It would be nice if this ad at least explained a little bit of the technical side of the solution.
It sounds vaguely like InnoDB’s concurrency control solution which uses tokens [0] as a unit of maximum work a query can perform.
0: https://dev.mysql.com/doc/refman/8.4/en/innodb-performance-t...